Actions after which I expect some kind of response seem to be more costly than the direct time cost they incur (for me, at least).
They also incur a decreased effectiveness of the time just afterward, due to causing me to be anticipatory – e.g. I am more likely to check up on whether I have received a response, and empirically find it generally harder to focus.
Examples of such actions
Messaging someone and awaiting a response
Leaving a comment or post on LessWrong or the EA forum
Being physically near people who might call upon me or talk to me
This one also makes “messaging someone and awaiting a response” more likely since I am more likely to need to coordinate in various ways with these people or follow up about things that were discussed in person
Ways to combat this cost
Go physically far away from people – e.g. plane, train, Airbnb, hotel, etc.
This seems to be a big part of why I am so productive on planes
Make a habit of asking “how important is this message, really?” before sending messages
If working in a space with others near, create a norm that only you can go into your working space, or that others should message you if they want to enter?
I am curious who else has or hasn’t noticed this kind of cost, and whether anyone has ideas for combatting it (my guess is that the policies I am currently operating under don’t respect this cost enough—e.g. I made 3-5x more progress toward my most important goal while on a plane yesterday than while not on a plane today, and it seems like if I had better mechanisms for appreciating the costs of being anticipatory, this difference could have been at least 10% smaller).
This is one of the reasons I prefer email and other long-form asynchronous communication methods over IMs and such.
This makes sense in the OS analogy. When you context-switch, you still have the previous context in cache. The question is: how long do you prioritize keeping the IM context in cache over bringing in the new context?
With batched communication, the answer is simple: drop the old context immediately and load the new context.
With IMs, you end up holding the old context in cache for a while in the hopes that you’ll context-switch back ‘soon’ and it’ll be useful. But of course this means you’re guessing as to how long it’ll be before you get a response… and if you don’t get a response meanwhile you’re running with less cache allocated to the new context.
It sounds like you could benefit from the “time management technique” of e-mail batching
Specifically, you set aside specific times each day when you will check your inbox/messages. Depending on your job role, that might be “every hour, on the hour” or “when I start work, get back from lunch, and right before I leave” or something in-between. Then after attending to anything there, you turn off your e-mail and other messaging applications, and do your productive work until the next check-in time.
Let the rest of your team know that asynchronous communications won’t be attended to immediately, and to call/use the one synchronous app if something needs immediate response (and do the same: don’t send an e-mail or Skype message if you need or expect or “anticipate” a reply within 5 minutes.